模拟植物生长算法的改进策略及桁架结构优化研究  

Improved strategies of plant growth simulation algorithm and optimization of truss structures

作　　者：石开荣[1,2] 阮智健 姜正荣[1,2] 张原[1,2] 林全攀 

机构地区：[1]华南理工大学土木与交通学院,广东广州510641 [2]华南理工大学亚热带建筑科学国家重点实验室,广东广州510641

出　　处：《建筑结构学报》2018年第1期120-128,共9页Journal of Building Structures

基　　金：亚热带建筑科学国家重点实验室开放课题(2012KB31);广州市科技计划项目(1563000257)

摘　　要：模拟植物生长算法（PGSA）是一种以植物向光性机理（形态素浓度理论）为启发准则的新型智能优化算法。对PGSA的基本原理进行了分析,指出并证实了该优化算法的局限性。在此基础上,提出了两种新的算法改进策略：形态素浓度计算的精英策略及智能变步长策略。前者通过在生长过程中快速剔除劣质生长点以提高算法的优化效率,后者通过不断变化步长以减少算法搜索时间。通过算例验证了所提出的改进策略可有效提高优化效率及解决算法缺乏终止判断机制的问题。最后采用改进PGSA和PGSA对典型桁架进行优化计算,结果表明改进PGSA的优化效率明显高于PGSA,在结构优化问题中具有较好的适用性。Plant growth simulation algorithm （PGSA） is one kind of intelligent algorithm which is based on the plant phototropism mechanism （morphaetin concentration theory）. The basic principle of PGSA was analyzed, and then the limitations of PGSA were pointed out and verified. According to these, two improved strategies were proposed： elite strategy of morphactin concentration＇ s calculation and strategy of intelligent variable step size. The former can remove the inferior growth points quickly during the growth process so as to improve the optimal efficiency, and the latter can reduce the search time by changing the step size. Through typical examples, both improved strategies have positive effect on optimizing efficiency and judgment mechanism of termination. In the end, through a truss optimization design problem, the result shows that improved PGSA has significantly higher optimization efficiency than PGSA. Improved PGSA has a good applicability in the structural optimization.

关 键 词：桁架 模拟植物生长算法 结构优化 精英策略 智能变步长策略 

分 类 号：TU323.4[建筑科学—结构工程]